NEW ERA OF BANKING PLATFORMS MIKHAIL KHASIN, SENIOR MANAGING - - PowerPoint PPT Presentation

NEW ERA 
 OF BANKING PLATFORMS

MIKHAIL KHASIN,

SENIOR MANAGING DIRECTOR & 
 HEAD OF CORE BANKING TRANSFORMATION PROGRAM SBERBANK

TECH NOLO

PLAT FORM

PARTNERS PARTNERS

REAL ESTATE B2C E- COMMERCE

LIFESTYLE

E-HEALTH

TELECOM

B 2 B S E R V I C E S B2B E- COMMERCE

Repair (furniture, 
 decor), other Purchase/lease 
 real estate Logistics

Materials, 


• ther

Cars and 
 equipment

Agriculture

Risk-management/
 ratings

Marketing, 


• ther

Business ops.

MNVO 
 Marketplace 


• f tariffs

Consulting P h a r m a c i e s

• Med. Institutions 


services Culture & leisure P a s s e n g e r 
 T r a n s p

• r

t a t i

• n

Media (incl. 
 social networks) Restaurants F

• d

,

• t

h e r Electronics, clothing

Trade financing

Payments

Loans

LCA

P 2 P T r a n s f e r s

NEW ERA 
 OF BANKING PLATFORMS

ANY ECOSYSTEM BASED 
 ON TECHNOLOGICAL PLATFORM

Hundreds millions 
 clients Petabytes 


• f data

Hundreds of thousands transactions per second

INNOVATIONS AS DRIVER FOR UNIQUE 
 PHENOMENON IN GLOBAL ECONOMY

2016 2015 2014 2013 2012 2011 2010 2009

Number of orders 
 per second Number of payments 
 per second Total amount of deals (yuan)

175 000

140 000 80 000 42 000 14 000 3 200 1 000 400

175 000

85 900 38 000 15 000 3 850 1 200 500 200

120,7 bln.

91,2 bln. 57,1 bln. 35 bln 19,1 bln. 5,31 bln. 1,94 bln. 590 mln.

NEW GENERATION OF BANKING PLATFORM

Analytics Area

Data Marts Analytical applications

Corporate memory (Hadoop)

Data Warehouse

Platform replica
 (archive of the platform) Detailed data model External Sources Replicas Models Реплики АC Банка Bank Systems Replicas

API

Client session data All active

• perations


All active products
 Archive

• f data to

a depth of at least 15 years

OMNI CHANNEL FRONT END BUSINESS-
 HUB PRODUCT
 FACTORIES DATA 
 FACTORIES

Client’s request Service Next Best Offer Information storing Client History Client Analytics Behavior Modeling Needs Forecast

Client 
 Profile EXTERNAL SITES

Activities in social networks

PUBLIC
 CLOUD

External Profile

EXTERNAL 
 ANALYTICS

COMPARISON OF THE TARGET SBERBANK PLATFORM 
 CLUSTER WITH THE LARGEST SUPERCOMPUTERS 
 OF THE WORLD*

Sberbank
 (Russia) Amazon Web Services
 (USA) National Center for Atmospheric Research (NCAR) (USA) Alibaba
 (China) MIT, Lincoln Laboratory
 (USA) Moscow State University - Research Computing Center
 (Russia) System Sberbank’s Platform Cluster Amazon EC2 C3 Instance cluster Cheyenne –
 SGI ICE XA Lenovo ThinkServer RD650 TX-Green - S7200AP Cluster Lomonosov 2 – 
 T-Platform A-Class Cluster Cores 56,000 26,496 144,900 84,000 41,472 42,688 Nodes 2,000 880 4,032

3,377 1,500

1,472 Theoretical Peak (Rpeak), TFlop/s 2,150 593.5 5,332.3 3,360 1,725.23 2,102 Memory, TB 1,536 103.5 198 218.75 121.5 92 * Data from worldwide TOP500 Supercomputer List (June 2017)

ARTIFICIAL INTELLIGENCE…

— it’s been a long time since it ceased to be a science fiction and has became something we carry in our own pockets daily

Apple’s Siri, Android’s Google Now, Yandex Alice, Personal Financial Assistants and other apps facilitate a brand new level of rendering information and financial services. Weekly the data-technology market brings new features enabling to propel AI even further across the industry. AI proved to be extremely sought after all the way from successful local business solutions to becoming a global financial trend as well as future banking cluster. Business models, processes, risks and experience are geared towards the general transformation wave.

BY 2020, THE MAJORITY OF NON-ROUTINE CAREER 
 PATHS WILL BE AFFECTED BY SMART MACHINES

“83% of professions paid less than $20 per hour will be taken by robots”.

Probability of automation 
 by a profession’s median hourly wage

— Council of Economic Advisors, USA

KEY TRENDS IN AI ENGAGING FOR BANKING

Chat bots Roboadvising Personalized offers Internet of Things Anti-fraud Operational efficiency

CHAT BOTS AND ROBO-ADVISING

Rendering information 


• n products & services

Provision of contact details Payment transaction posting Financial advising for clients

AI in banks. Key trends (1/6)

ROBO-ADVISORS AS A PROMISING AI APPLICATION: 
 CASE STUDY

Robo-advising has become an alternative financial consulting

service provider on banking issues as well as specific purchases and other monetary on-line transactions.

Robo-advisors offer substantial advantages in on-line trading.

First and foremost, this is due to single-click applications, account creation in a real time mode, monitoring, latest news and ability to process multiple deals at once. The brokers disseminated across social media improve data accessibility and comprehensiveness, and make communication with clients to be more targeted and easy job.

Automation enables to provide information in 24/7 mode in a

less costly manner. Robo-advisors can be made accessible either via your desktop or as a mobile app acting as portfolio managers that are capable to identify risks and devise streamlined investment strategy.

AI in banks. Key trends (2/6)

Estimated U.S. Robo-advisors assets 
 under management


($ trillions)

Growth due to invested assets 
 (cash, bank deposits) Growth due to non-invested assets 
 (Credit risk instruments, stock and mutual funds)

Source : A.T.Kearney simulation model

2016E 2017E 2018E 2019E 2020E

PERSONALIZED OFFERS AND IMPROVED LOYALTY

Recommending banking products and purchases

(loyalty programs by different retailers) inter alia – relying upon client’s info from social media

Identifying the existing client’s B2B network and

providing recommendations on engaging with new counterparties

Simulating financial risks for small businesses

(default, cash deficiency etc.) in a real time mode; recommending new target strategies and products AI in banks. Key trends (3/6)

IOT (INTERNET OF THINGS)

Management and tracking of the leased assets Smart insurance services for retail clients (health

coverage, auto-loans etc.)

Smart Home + Daily Shopping: means ordering, public

utility bills payment, TV content subscription

Banking of Things: transfer the payments function from

people to devices (e.g., cars pay for gas, parking and using of toll roads)

AI in banks. Key trends (4/6) It is expected that the number of IoT-connections will grow by 23% annually within 2015 to 2021. IoT devices will encompass more than 16 billion out of 28 billion connected objects by the end of the projected period.

ANTI-FRAUD. INSIDE AND OUTSIDE THREATS

Attributes hinting that a credit card is used by an

authorized person

Attributes of so called “droppers” identified based on

specificity of credits and transactions via online-bank and ATMS

Identifying fraudulent salary projects (loans, cash-pull) Identifying unauthorized debit transactions from client’s

accounts and cards

Errors in parameterization of the Bonus programs on credit

cards resulting in unjustified mark-ups and loss & damage

Cash-pull schemes, including via online-bank and credit

cards

Abuse in conversion transactions for both retail and

corporate entities

Unauthorized connections of online-bank to client’s

accounts 
 and credit cards issuing without the knowledge of the client

Unauthorized limit increase for credit cards

AI in banks. Key trends (5/6)

OPERATIONAL EFFICIENCY

Identifying deviations in transaction execution and

automatic correction hereof

Natural Language Processing – algorithms for analysis

and generation of legal claims

Monitoring and prediction of infrastructure failure

(ATMs, IT-resources)

Streamlining of cash flow cycle in cash departments

and ATMs. Streamlining of collection services

Optimization of recruitment and hiring processes (CV

review and initial screening)

Speech analytics in a real time mode for call centres

and branches (consultation quality management)

AI in banks. Key trends (6/6)

MACHINE LEARNING MECHANISMS

Identification of bottlenecks in transaction

processing

Identification of root causes behind exceptions

that occur upon documents execution and categorization thereof

Identification of major user’s mistakes in the

system

Analysis of the system users & clients activities.

Predicting the load on the platform

Analysis of client's product preferences and

anticipating future actions

Best personal offer